Yeah, tailwind JIT traded performance for less concise editor tooling (which was a good trade, if you ask me). Not sure if there’s going to be a fix for that in the future due to the way how extraction works now. If there is, it’s probably going to happen in their editor extension (which imo is a better place for doing class name verification).
Suggest the verification pass to the Tailwind project
It seems wrong to me that the onus of a verification check is to be placed on every possible project out there rather than having it solved once and for all by the upstream. It is doubly true with on-demand building of the CSS, since it won’t be possible for a downstream tool to solve the problem.